Self-organization in the spelling of English suffixes: The emergence of culture out of anarchy* Kristian Berg, University of Oldenburg Mark Aronoff, Stony Brook University COMMENTS WELCOME *This paper was supported by the German Science Foundation (DFG) project “Prinzipien der Wortschreibung im Deutschen und Englischen” (‘Principles of word spelling in German and English’) and a DAAD post-doctoral scholarship to Kristian Berg held at Stony Brook. 1 1 Introduction English culture is famously unprincipled. For starters, England has never had a constitution. Magna Carta, signed eight hundred years ago, is sometimes held up as an English bill of rights, but the only ones whose rights it protected were barons. Since then and before then, England and the rest of Great Britain have gotten along very well without a constitution of any sort, thank you, making the country a very conspicuous exception among the constitutionally based nations that dominate the modern world. Even when it comes to simple written law, the English are sorely lacking. Not just England, but almost all of its former colonies stand out in following common law, which is governed largely by precedent rather than by written statute. No constitution, no laws, what kind of culture is that? No spelling or grammar rules either. The English language has just as staunchly and just as successfully resisted statutory regulation as has its government and its legal system. The founding of the Accademia della Crusca in Florence in 1582 and the Académie Française in 1735, both devoted to overseeing language, led such British scientific and literary luminaries as Robert Boyle, Jonathan Swift, and Joseph Addison to propose a similar governing body for the English language. Strongly opposed by Samuel Johnson on the grounds of “English liberty,” (Martin 2008, p. 197), the idea quickly fell out of fashion, leaving the language with no government or police.1 Nonetheless, as we will show, just like a spoken language, English spelling has arrived at a system despite the lack of any overt guidance.2 It has been suggested that the English language has overwhelmed the globe because it has no one to police it. No nation or authority of any sort owns English and neither the United Kingdom nor the United States has an official national language. This permits anyone in the world to use English as they will, making up their own words and constructions with no official interference. There is not even an authority anywhere in the world governing how English is spelled. Present-day English spelling varies from country to country, enforced only by local editorial practice, which may differ from one publisher or organization to another. English spelling, in any of its current incarnations, appears to be as lawless as it is ungoverned, anarchy run amok. It is notoriously unphonetic, rivaled in that regard only by French spelling, and examples abound of the same sound spelled in different 1A British Academy was eventually chartered by royal decree in 1902. According to its own official history, the academy “was first proposed in 1899 in order that Britain could be represented at a meeting of European and American academies” (because it had none!). This academy, however, has never had any jurisdiction over the English language. 2 The analogy between English spelling and English common law is not entirely apt. Common law rests on court precedent (stare decisis), which depends on having access to the recorded decisions of individual judges within a legal hierarchy. For spelling we have neither courts nor records to guide us; one aspect of the analogy that does hold is that we can only understand the current system through its history. 2 ways. The major consistency is lexical, in the spelling of individual words: a given word will be spelled in one way within a given tradition. Word spellings may differ from one tradition to another, sometimes in complex ways: US spelling uses judgment while the usual British spelling is judgement, except that the British use judgment to specifically denote a judicial decision. homophones that are not also homographs, distinct words that sound the same but are spelled differently, provide a striking example of lexical consistency. Each of the three identical-sounding words pare, pair, and pear has a consistent but distinct spelling. Among alphabetic writing systems, the use of distinct spellings to differentiate words visually is most pronounced in English and French. The origins of this strategy are unclear, though both systems were fixed at about the same time, in the half-century or so after 1650. The downside of the strategy is that it wreaks havoc with sound-spelling correspondences, as most critics of French and English spelling have noted. This article is a study of the middle ground between the spelling of sounds and the spelling of words: the spelling of affixes, specifically suffixes. We will show that the spelling of any given English suffix is quite consistent, despite the absence of any external authority making it so. This observation is not entirely new. It has often been remarked (Carney 1994: 18ff.) that the two most common English inflectional affixes, <-s> and <-ed>, while they vary in form depending on their phonological environments, do not vary in spelling. <-s> can be pronounced as either [s], [əz], or the default [z], depending on the sound at the end of the word to which it is attached, while <-ed> is either [t], [əd], or [d]. In both cases the spelling remains the same despite the different pronunciations (which are admittedly predictable). The suffix <-s> is polyfunctional. It can represent either the plural of nouns (cats), the third person singular present of verbs (tends), or the possessive of nouns (men’s), but the possessive is orthographically distinct, carrying an apostrophe before it, so that the plural cats and the possessive cat’s differ from one another in form, though the plural noun dogs and the singular verb dogs are indistinguishable.3 Chomsky and halle (1968) have famously noted that lexical consistency extends even to derived words, citing such sets as {sane, sanity}, {sign, signify}, and {electric, electricity, electrician}. In each case, the spelling of the base word remains the same throughout, despite the phonological changes consequent on suffixation. Although they both signal some sort of constant, there is a subtle difference between the constant spelling of lexemic stems despite differences in pronunciation in cases like electric/electricity/electrician and the constant spelling of suffixes that is the object of our study here. The lexemes are spelled in the same way despite differences in pronunciation in different environments. A given suffix is spelled the same way across different words that contain it and this constant spelling differs from that of the same phoneme sequence in instance where this phoneme sequence does not represent the suffix. 3 The regular genitive plural marker is <-‘>, a silent apostrophe, since the genitive suffix does not occur after the plural [-s] on account of haplology. This may be the only case of a true zero marker in English spelling. 3 Put another way, the spelling system follows the general pattern of distinguishing homophones, peculiar to English and French spelling, which we saw already in examples like pair/pear/pare, but it extends the pattern beyond words to word endings. As mentioned above and as we will show, affixes are spelled differently from homophonous sequences that that happen to fall at the ends of lexical words. The system spells the denominal adjectival suffix <ous> consistently, while all other words that end in the same sequence are spelled differently. The words nervous, office, and tennis all end in the phonological sequence [̌əs] but only nervous contains the suffix. We have arrived at the two larger questions that we address here. First, to what extent does present-day English spelling call attention to the spelling of individual affixes beyond the two inflectional suffixes? Second, how did it reach its current state? The answer to the second question bears on the larger and much more interesting general question of how a system can emerge in the absence of any stated principles or guiding hand, how English spelling, like English common law, came to take on the shape that it has today. We will show that affixal constancy is characteristic of the current state of English spelling. But matters were not always so. Overall, our study reveals that Matthew Arnold (1869) was wrong in contrasting culture and anarchy. Culture, at least in this case, arose out of anarchy. From a much wider perspective, the emergence of systematicity in English spelling is another example of the workings of competition, the struggle for existence. English spelling from the Middle English period through the end of the seventeenth century was unsettled because history had provided numerous ways to spell the same word. Queen Elizabeth, for example, who left a large legacy of autograph correspondence, had a single spelling for only half the words that she used in the documents that have been preserved (707 out of 1389), though she was not entirely unsystematic: 523 of the remaining 682 words showed only two variants (Evans 2012). The history of English spelling since 1600 shows what happened to many of the available variants: their distribution became lexically fixed, for both lexemes and suffixes. In ecological terms, each variant spelling has found its niche. What defines these niches, though, is not spelling. Spelling simply fills niches that are made possible by the morphology of the language. As one of us once noted, “written language is a product of linguistic awareness, the objectification of spoken language. Any orthography must therefore involve a linguistic theory,” albeit an implicit one (Aronoff 1985, p.